0521853036 cambridge university press style language variation and identity aug 2007

Coupland draws onand integrates a wide variety of contemporary sociolinguisticresearch as well as his own extensive research in this field.The emphasis is on how social meanings are made

Style: Language Variation and Identity

Style refers to ways of speaking – how speakers use theresource of language variation to make meaning in socialencounters This book develops a coherent theoreticalapproach to style in sociolinguistics, illustrated with copiousexamples It explains how speakers project different socialidentities and create different social relationships throughtheir style choices, and how speech-style and social contextinter-relate Style therefore refers to the wide range of stra-tegic actions and performances that speakers engage in, to con-struct themselves and their social lives Coupland draws onand integrates a wide variety of contemporary sociolinguisticresearch as well as his own extensive research in this field.The emphasis is on how social meanings are made locally, inspecific relationships, genres, groups and cultures, and onstudying language variation as part of the analysis of spokendiscourse

N I K O L A S C O U P L A N D is Professor and Research Director

of the Cardiff University Centre for Language and cation Research He is a founding co-editor of the Journal

Communi-of Sociolinguistics

Already published in the series:

Politeness by Richard J Watts

Language Policy by Bernard Spolsky

Discourse by Jan Blommaert

Analyzing Sociolinguistic Variation by Sali A Tagliamonte

Language and Ethnicity by Carmen Fought

Forthcoming titles:

World Englishes by Rakesh Bhatt and Rajend Mesthrie

Bilingual Talk by Peter Auer

Language Variation and Identity

N I K O L A S C O U P L A N D

Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São PauloCambridge University Press

The Edinburgh Building, Cambridge CB2 8RU, UK

First published in print format

ISBN-13 978-0-521-85303-3

ISBN-13 978-0-511-35005-4

© Nikolas Coupland 2007

2007

Information on this title: www.cambridge.org/9780521853033

This publication is in copyright Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press

ISBN-10 0-511-35005-8

ISBN-10 0-521-85303-6

Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate

Published in the United States of America by Cambridge University Press, New York

www.cambridge.org

hardback

eBook (NetLibrary)eBook (NetLibrary)hardback

List of figures and tables vii

Preface and acknowledgements ix

Transcription conventions xiii

2.2 Limits of the stratification model for style 37

2.3 ‘Standard’ and ‘non-standard’ 42

2.4 ‘Non-standard’ speech as ‘deviation’ 45

2.5 Social structure and social practice 47

3 Style for audiences 54

3.1 Talking heads versus social interaction 54

3.2 Audience design 58

3.4 Some studies of audience design and speech

3.5 Limits of audience-focused perspectives 74

4 Sociolinguistic resources for styling 82

4.1 Speech repertoires 82

4.2 The ideological basis of variation 85

4.3 Habitus and semantic style 89

4.4 Language attitudes and meanings for variation 93

v

4.5 Metalanguage, critical distance and performativity 99

4.6 Sociolinguistic resources? 103

5 Styling social identities 106

5.1 Social identity, culture and discourse 106

5.2 Acts of identity 108

5.3 Identity contextualisation processes 111

5.4 Framing social class in the travel agency 115

6 High performance and identity stylisation 146

6.1 Theorising high performance 146

6.2 Stylisation 149

6.3 Decontextualisation 155

6.4 Voicing political antagonism – Nye 156

6.5 Drag and cross-dressing performances 163

6.6 Exposed dialects 171

7 Coda: Style and social reality 177

7.1 Change within change 177

7.2 The authentic speaker 180

7.3 The media(tisa)tion of style 184

References 189

Index 206

Figures and tables

Figures:

The International Phonetic Alphabet

Consonants (Pulmonic)

Vowels xiv

Figure 2.1: Class and style stratification for (th) 33

Figure 2.2: Class and style stratification for (r) 34

Figure 2.3: Distributions of variants of (e), (ay) and (wedge)

among jocks and burnouts, boys and girls 52

Figure 3.1: Percentage of intervocalic /t/ voicing by four

newsreaders on two New Zealand radio stations,

Figure 3.2: Sue’s convergence on (intervocalic t) voicing to

five occupation classes of client; input level taken asSue’s speech to ‘her own class’ 73

Tables:

Table 3.1: Foxy Boston’s vernacular usage in Interviews III and IV 66

Table 3.2: Percentages of less ‘standard’ variants of five

sociolinguistic variables in four ‘contexts’ of

Sue’s travel agency talk 72

Table 4.1: Mean ratings (whole sample, 5,010 informants) of

34 accents of English according to social

attractiveness and prestige 98

Table 6.1: Phonetic variables generally distinguishing South

Wales Valleys English and Received Pronunciation 158

vii

Preface and acknowledgements

In the new world of sociolinguistics, the simple concept of ‘style’ has alot of work to do The idea of ‘stylistic variation’ emerged fromWilliam Labov’s seminal research on urban speech variation andlanguage change, and it existed there in order to make a few keypoints only As Labov showed, when we survey how speech varies,

we find variation ‘within the individual speaker’ across contexts oftalk, as well as between individuals and groups Also, when individualpeople shift their ways of speaking, survey designs suggested that they

do it, on the whole, in predictable ways that are amenable to socialexplanation

From this initially narrow perspective, crucial as it was in ing a basic agenda, a sociolinguistics of style has steadily come toprominence as a wide field of research, whether or not researchersuse the term ‘style’ to describe their enterprise Style used to be amarginal concern in variationist sociolinguistics Nowadays it points

establish-to many of the most challenging aspects of linguistic variation, inquestions like these: How does sociolinguistic variation interfacewith other dimensions of meaning-making in discourse? What stylis-tic work does variation do for social actors, and how does it blend intowider discursive and socio-cultural processes? Are there new valuesfor variation and for style in the late-modern world?

When we work through issues like these, some important ries shift For one thing, the study of sociolinguistic variation becomesvery much wider The canonical study of language variation andchange will always remain a pillar of sociolinguistics, but it need not

bounda-be an autonomous paradigm One of my ambitions for the book is toshow what variation study is like when it ‘goes non-autonomous’ Theboundary between ‘dialect variation’ and the social construction ofmeaning in discourse starts to collapse Theories and sensitivitiesfrom different parts of sociolinguistics start to coalesce – interactionalsociolinguistics, pragmatics, anthropological linguistics and even

ix

conversation analysis do not need to stand outside of variationism,nor it outside them.

My own thinking on sociolinguistic style has spanned two-and-a-halfdecades, although it remains to be seen whether this particular quanti-tative index (like some other quantitative measures that come up forreview in the book) makes a meaningful difference I was enthused towrite this book mainly because of the acceleration of sociolinguisticinterest in things ‘stylistic’ and ‘contextual’ and ‘socially meaningful’ inthe last decade, prompted by some remarkable new waves of research Iwon’t attempt to list the relevant names and paradigms here – they fillout the pages of the book But I would like to make a few biographicalnotes, by way of personal acknowledgement

I had begun writing about style in the late 1970s, when the themeemerged from my doctoral research on sociolinguistic variation inCardiff, the capital city of Wales I was fortunate to start long-runningdialogues, soon after that, with Allan Bell and Howard Giles In theirown research they developed new relational perspectives on spokenlanguage variation that opened up an entirely new theoretical chapterfor sociolinguistics I continued to collaborate with Howard Giles overmany years on various themes that lay at the interface between socio-linguistics and social psychology I have been fortunate to be able todevelop some of that work, more recently, in collaboration with PeterGarrett and Angie Williams in Cardiff, and more recently still withHywel Bishop

After some scratchy ink and pen exchanges about his evolving theory

of audience design in the very early 1980s, Allan Bell and I maintainedclose links, latterly in co-editing the Journal of Sociolinguistics That partic-ular collaboration ensured we would have no time to write collabora-tively about style, although we had firmly intended to do this I have nodoubt that this book would have been much the better if Allan and I hadachieved our aim of writing a similar book together

As the Centre for Language and Communication Research at CardiffUniversity grew and diversified through the 1980s and 1990s, several

of my colleagues there were involved in developing new tic fields, particularly critical and interactional approaches to lan-guage and society The study of style needed the sorts of insight thatthey were developing in their own and in our joint research In partic-ular there has been the formative effect of my many collaborationswith Adam Jaworski, for example on metalanguage, sociolinguistictheory and discourse analysis My other Cardiff colleagues, includingTheo van Leeuwen and Joanna Thornborrow, have again been impor-tant sources of inspiration My research collaborations with Justine

Coupland, for example on the theme of discourse and ideology, socialidentities in later life and on relational talk, have been where I devel-oped most of the ideas behind the present book, although her contri-butions to this book are far too pervasive to summarise.

Apart from those already mentioned, a long list of people have madevery valuable input into my thinking and writing about ‘style’, whetherthey recall it or not No doubt with unintended omissions, let me thankPeter Auer, Mary Bucholtz, Janet Cotterill, Penelelope Eckert, AntheaFraser Gupta, Janet Holmes, Tore Kristiansen, Ben Rampton and JohnRickford Thanks also to Rachel Muntz and Faith Mowbray for theirhelp in connection with the BBC Voices research that has a walk-on part

in Chapter4 Reading groups convened by Julia Snell, Emma Moore andSally Johnson fed back some valuable criticisms on parts of the text AyoBanji made extremely helpful input into compiling the Index AllanBell, Adam Jaworski and Natalie Schilling-Estes, as well as RajendMesthrie, read and commented on the whole manuscript in draftform, for which I am extremely grateful

I have summarised and rewritten parts of my previously publishedwriting in this book The main sources in this connection, listed in theReferences section, are Coupland 1980, 1984, 1985, 1988, 2000b,

2001b,2001c,2003, in pressa, in pressb, Coupland and Bishop2007,Coupland, Garrett and Williams2005, Coupland and Jaworski2004 I

am particularly grateful to my co-authors for letting me rework someparts of this material here Figures 2.1 and 2.2 are adapted fromFigures 7.23and 7.11in Labov (2006 )

The disciplinary boundary-shifting that I referred to above haspresented me with the problem of knowing where to draw the linearound style in this book I have given most space to those studies ofhow classical forms of sociolinguistic variation – what most peoplecall accent and dialect features – are worked into discursive socialaction and where they make meaning at the level of relationships andpersonal or social identities As I say later, this is a rather artificialboundary to try to police, because my motivating concerns for thebook are social meaning and social identity, much more than socio-linguistic variation itself For example, I would have liked to includesome detail on the discursive management of age-identities in laterlife (an area of my own my research with Justine Coupland) But thiswould have taken the book away from indexical meanings linked tothe domains of social class, gender and racial/ ethnic identities, which

is where style research has been most active to date

This book can be read as a critique of variationist sociolinguistics.Meaning-making through talk has not been what variationists have

generally tried to explain, although it has seemed to me a strangeomission It is all the more strange when we think of William Labov’scommitment to the politics of language variation, his interest from theoutset in the social evaluation of varieties, and his ground-breakingwork in narrative analysis and interactional ritual His followers in thefield of variationist sociolinguistics have not often been able to main-tain that breadth In order to bridge back into questions of socialmeaning, I have found it important to challenge some of the assump-tions of variationist research These are mainly its dogged reliance onstatic social categories, its imputation of identity-values to numericalpatterns (quantitative representations of linguistic variation), and itsthin account of social contextualisation.

I fully recognise that, and celebrate the fact that, variationist linguistics has taken great strides through keeping within these con-straints, when research questions have been formulated at the level oflinguistic systems and how they change But I think we need a socio-linguistics of variation for people and for society, as well as (not insteadof) a sociolinguistics of variation for language ‘Sociolinguistic style’ hasbeen the rubric under which quite a lot of that extension of theprogramme has already been achieved, and where further progress

socio-is clearly in prospect ‘Stylsocio-istics’, as a label for a sub-dsocio-iscipline oflinguistics, has a dated feel to it, and so does ‘style’ But in the context

of sociolinguistics, style nevertheless points us to a range of highlycontemporary phenomena We seem to find meaning in our livesnowadays less through the social structures into which we havebeen socialised, and more through how we deploy and make meaningout of those inherited resources How social reality is creatively styled

is a key sociolinguistic question, and the main question in whatfollows

NCJuly 2006

Transcription conventions

Where necessary, International Phonetic Alphabet (IPA) symbols areused to identify consonant and vowel qualities, as in the followingcharts (as shown over)

T H E I N T E R N A T I O N A L P H O N E T I C A L P H A B E T

E X T R A C T S O F T R A N S C R I B E D C O N V E R S A T I O N

These are numbered consecutively within each chapter Where ble, I have re-transcribed data extracts from the original sources in theinterests of simplicity and consistency Wherever possible, these tran-scriptions use orthographic conventions, but with the following addi-tions and deviations:

[quietly] stage directions and comments on context or spoken

delivery

speech, showing beginning and end points of overlap

(( )) inaudible speech sequence or unreliable transcriptionitalics sequences of particular analytic interest, explained in

the text

Any other conventions used in particular extracts are explained inthe text

xiii

1 Introduction

1 1 L O C A T I N G ‘ S T Y L E ’

‘Style’ refers to a way of doing something Think of architectural stylesand the striking rustic style of house-building in rural Sweden Thatparticular style – what allows us to call it a style – is an assemblage ofdesign choices It involves the use of timber frames, a distinctivelytiered roofline, a red cedar wood stain and so on We can place thisstyle It belongs somewhere, even if the style is lifted out of its hometerritory and used somewhere else It has a social meaning The same

is true for styles in all other life-domains Cultural resonances of time,place and people attach to styles of dress and personal appearance

in general, to styles in the making of material goods, to styles of socialand institutional practice, perhaps even to styles of thinking Wecould use David Machin and Theo van Leeuwen’s (2005) idea of ‘socialstyle’ to cover all these The world is full of social styles

Part of our social competence is being able to understand theseindexical links – how a style marks out or indexes a social difference –and to read their meanings The irony is that, if we ourselves areclosely embedded in a particular social style, we may not recognisethat style’s distinctiveness Reading the meaning of a style is inher-ently a contrastive exercise You have to find those red cedar buildings

‘different’ in order to see them as having some stylistic significance.This is the old principle of meaning depending on some sort of choicebeing available But style isn’t difference alone When we use the term

‘style’ we are usually attending to some aesthetic dimension of ence Styles involve a degree of crafting, and this is why the word

differ-‘style’ leaks into expressions like ‘having style’, ‘being in style’ or

‘being stylish’ The aesthetic qualities of styles relate, as in the case

of the Swedish red cedar buildings, to a process of design, howevernaturalised that process and its results might have become in ourexperience We talk about ‘style’ rather than ‘difference’ when we

1

are aware of some holistic properties of a practice or its product Astyle will ‘hang together’ in some coherent manner Engagement withstyle and styles, both in production and reception, will usually imply acertain interpretive depth and complexity Although we are consider-ing ‘style’ as a noun at this point, when we refer to ‘a style’ and to

‘styles’ (plural), and giving styles a quality of ‘thing-ness’, the idea ofstyle demands more of a process perspective I think we are mainlyinterested in styles (noun) for how they have come to be and for howpeople ‘style’ (verb) meaning into the social world ‘Styling’ – theactivation of stylistic meaning – therefore becomes an importantconcept in this book

This general account of style can of course be applied to linguisticforms and processes too We are all familiar with the idea of linguisticstyle, and most people will think first of language in literary style.Literary style relates to the crafting of linguistic text in literary genresand to an aesthetic interpretation of text This book is about style inspeech and about ways of speaking, not about literary style, although itwould be wrong to force these areas of study too far apart The book isabout style in the specific research context of sociolinguistics, whereconcepts very similar to ‘social style’ have been established for severaldecades The general sociolinguistic term used to refer to ways ofspeaking that are indexically linked to social groups, times and places

is dialects Dialects are social styles Some dialects are in fact rather likered cedar timber buildings, redolent with meaningful associations ofrurality and linked to particular geographical places They have strongcultural associations, especially when we look at them contrastively.Dialectologists have traditionally looked for boundaries between dia-lect regions, and traced the evolution of dialects over time and theconsequences of dialects coming into contact with each other(Chambers and Trudgill1999)

We are likely to think of dialects in this sense as being the socialstyles of yesteryear, largely out of step with the social circumstances ofcontemporary life But dialect differences are of course a character-istic of modern life too Dialects are evolving social styles and they can

be read for their contemporary as well as their historical associations –associations with particular places (geographical dialects) and withparticular social groups (social dialects) Dramas associated with dia-lect are played out as much in cities as in rural enclaves, and socio-linguistics for several decades has enthusiastically teased out thecomplexities of language variation in urban settings The humanand linguistic density of cities invites an analysis in terms of ‘struc-tured difference’ Cities challenge the view that one discrete social

style (e.g a dialect) is associated with one place, which was the basicassumption in the analysis of rural dialects It has become the norm toconsider cities as sociolinguistic systems that organise linguistic var-iation in complex ways But understanding the social structuring ofstyles, even in the sophisticated manner of urban sociolinguistics, isnot enough in itself We need to understand how people use or enact orperform social styles for a range of symbolic purposes Social styles(including dialect styles) are a resource for people to make manydifferent sorts of personal and interpersonal meaning As I suggestedmight be generally true for intellectual interest in style, what mattersfor linguistic style is more to do with process than with product, more

to do with use than with structure Stylistic analysis is the analysis ofhow style resources are put to work creatively Analysing linguisticstyle again needs to include an aesthetic dimension It is to do withdesigns in talk and the fashioning and understanding of socialmeanings

So this is not a book about dialectology either My starting point iscertainly the sociolinguistics of dialect, as it has been carried forward

by variationist sociolinguistics in the tradition of William Labov’sresearch This is where the term ‘style’ was first used in sociolinguis-tics, and one of my aims for the book is to map out the main stepsthat sociolinguists have taken using the concept of style This willinitially be a critical review, focusing on the limited horizons of styleresearch in variationist sociolinguistics The positive case to be made,however, is that, under the general rubric of style, sociolinguistics canand should move on from the documenting of social styles or dialectsthemselves It should incorporate the priorities I have just sketched –analysing the creative, design-oriented processes through whichsocial styles are activated in talk and, in that process, remade orreshaped This means focusing on particular moments and contexts

of speaking where people use social styles as resources for making It means adding a more active and verbal dimension (‘stylingsocial meaning’) to sociolinguistic accounts of dialect (‘describingsocial styles’)

meaning-To set the scene for later arguments and debates, several core cepts need to be explored in this introductory chapter First we need toconsider variationist sociolinguistics and its general approach to style.Then we will look back at the early history of stylistics (the generalfield of research on style in linguistics), to appreciate the climate inwhich sociolinguistics first came to the idea of style The idea of socialmeaning then comes up for initial scrutiny Looking ahead to themore contemporary research that this book mainly deals with, we

will then consider research methods and the sorts of sociolinguisticdata that we can deal with under the heading of style research Thewider relevance of style to contemporary social life, which can becharacterised by the term ‘late-modernity’, is then reviewed Finally

in this chapter, I give a short preview of the structure of the rest ofthe book

1 2 V A R I A T I O N I S M I N S O C I O L I N G U I S T I C S

Sociolinguistics is, as they say, a broad church The blander definitions

of sociolinguistics refer to studying language ‘in society’ or language

‘in its social context’ Other definitions focus on studying linguisticdiversity or language variation What these simple definitions have incommon is that they give priority to language, then add some sum-mary idea of what aspect of language is to be given priority (itsvariability) or what sort of data is to be given priority (social manifes-tations of language) Definitions like these have to be understoodhistorically It was once important to stress ‘social contexts’ in defin-ing sociolinguistic priorities in order to challenge types of linguisticswhere actual occurrences of spoken language were not given priority.Even though most people would agree that using language is aninherently social process, sociolinguists needed to make a case forobserving language as it is used in everyday life and for not relying onintuited or fabricated instances of language Stressing variability hasbeen important in order to resist the ideological assumption that whatmatters in language is linguistic uniformity and ‘standardness’.William Labov used the notion of secular linguistics to describe hisapproach to language variation and change The idea was that study-ing variable language forms, ‘non-standard’ as well as ‘standard’forms, challenges what we might think of as the high priesthood oftheoretical linguistics and its reliance on idealised linguistic data Italso challenges the belief that ‘standard’ language is more orderly andmore worthwhile than ‘non-standard’ language

But the study of language variation and change has been in themainstream of sociolinguistics for four decades Variationist sociolin-guistics, as the approach developed by Labov is generally called,has developed its own powerful principles of theory and method(Chambers 1995/2003; Labov 1966, 1972a, 1972b, 1994, 2001a;Chambers, Trudgill and Schilling-Estes2004) In this book I intend totake the considerable achievements of variationist sociolinguistics forgranted, and to ask what it has not achieved, particularly in relation to

the notion of style and the active dimension of styling So, as I havementioned, my orientation is a critical one, although I intend it to beconstructively so The negative part of my argument is that variation-ist sociolinguistics has worked with a limited idea of social context –and styling is precisely the contextualisation of social styles The surveydesigns of variationist research, which have been remarkably success-ful in revealing broad patterns of linguistic diversity and change, havenot encouraged us to understand what people meaningfully achievethrough linguistic variation Variationist sociolinguistics has pro-duced impressive descriptions of social styles, but without affordingmuch priority to contextual styling.

What then are the general features of the variationist approach?Sociolinguistic surveys of language variation give us detailed descrip-tions of how linguistic details of regional and social accents anddialects are distributed (‘Dialect’ is a general term for socially andgeographically linked speech variation, and ‘accent’ refers to pronun-ciation aspects of dialect.) Speakers are not fully consistent in howthey use accent or dialect features Their speech will often, for exam-ple, show a mixture of ‘standard’ and ‘non-standard’ forms of the samespeech feature Nor are individuals within any particular social cate-gory identical in their speech So the sort of truth generated in varia-tionist research is necessarily one based in generalisations andstatistical tendencies These are ‘probabilistic’ truths, expressingdegrees of relative similarity and dissimilarity within and acrossgroups of speakers and social situations The convention is to produceaveraged statistical values (e.g percentages of people’s use of a partic-ular linguistic feature in a particular social situation, or factor load-ings in statistical tests) to represent patterns of linguistic variation So,accent variation between two different groups of speakers is usuallyrepresented as the difference between one statistical value (perhaps apercentage) and another

Variationist research has very expertly shown that ‘speaking ently’ has to be defined in several stages Stage one is typically toidentify a group of people who share a geographical characteristic,such as living in the Midlands city of Birmingham in England, or forthat matter Birmingham in Alabama in the Southern USA Within thisterritory or ‘community’ of people who have lived in the city for all ormost of their lives, sub-groups are identified based on social criteria.This sort of classification isolates, to take a random example, thecategory of ‘young females in Birmingham with working-class jobs’,distinguishing them from other social categories In a second stage, theresearch samples the speech of the different groups, usually through

extended one-to-one sociolinguistic interviews The researcher thencounts how often a particular speech feature is used.

For example, in the English Birmingham, the issue might be howoften each speaker pronounces the diphthong vowel in words likeright and time with a phonetically backed and rounded starting point

In this example, the local Birmingham pronunciation [OI] is in sition to [aI] which is the less localised and more ‘standard’ variant inEngland Phonetic forms occupying intermediate positions betweenthese variants might also be recognised Variant forms of sociolinguis-tic variables tend to be influenced by the details of their linguisticplacement For pronunciation variables (linked specifically to a spea-ker’s accent, then), the positions that different pronunciation formsoccupy in the stream of speech-sounds, and the sets of words that theyoccur in, are factors that are likely to impact on the frequency withwhich they are used These patterns might affect everyone’s speech Atypical finding would then be that most speakers in the sample would

oppo-in fact use a mixture of different pronunciation forms – e.g usoppo-ing both

‘standard’ and ‘non-standard’ variants of this sociolinguistic variable(ai) But overall frequencies of use would very probably differ acrossspeakers and sub-groups when statistical averages are taken

At the end of the process of categorising and counting the tion of various linguistic variants in a body of data, a type of statis-tical truth would emerge It might allow us to say that, overall,Birmingham speech does indeed have some distinctive tendencies ofpronunciation – different from the speech of other regions and from

distribu-‘standard English’ pronunciation That is, descriptively speaking,Birmingham speech is a relatively distinctive social style The descrip-tive evidence would go some way towards distinguishing the city as a

‘speech community’, even though the ‘standard’, less-localised forms

of speech crop up in Birmingham too But people living outside thecity would use some of the local or ‘non-standard’ feature less oftenthan those living in the city, or not at all Looking at how speech issocially organised within the city, we would probably be able to saythat the speech of particular social sub-groups in Birmingham differs

in some statistical respects Perhaps, overall, women in Birminghamuse the [OI] feature in words like right and time less often than men do.Perhaps women with more prestigious jobs use it less than womenwith low-prestige jobs So there are social styles, at least in a quanti-tative sense, associated with these groups too

Labov, however, doesn’t use the term ‘style’ in this sense He refers

to what I am calling ‘social styles’ of speech simply as ‘social tion’ He reserves the terms ‘style’ and ‘stylistic variation’ for a further

sort of language variation that can be detected in sociolinguistic views (e.g Labov1972b) This is when he is able to show that, again in astatistical sense, individual people speak ‘less carefully’ at somepoints in an interview than they do at other points When they arebeing ‘less careful’ or more relaxed they will typically use features ofthe local style more frequently than in their supposedly normal inter-view speech In this way Labov introduced the idea of ‘stylistic varia-tion’ to refer to ‘intra-individual’ speech variation – variation ‘withinthe speech of single individuals’ This became a very familiar claim incommunity-based studies of language variation and change, and wewill look at it in much more detail in Chapter2 But it is important tonote that, although Labov is mainly concerned with social style at acommunity level, his original insight about stylistic processes related

inter-to the individual speaker and inter-to particular social contexts of speaking.That is, he was interested in what happens when an individualspeaker delivers a version of a social style in a range of particularspeaking situations This proves to have been a seminal insight As weshall see, however, the survey methods that Labov pioneered tend not

to give priority to the local processes through which this happens.They orient much more to styles than they do to styling The conven-tion of basing variationist research on speech in interviews clearlylimits the range of social contexts in which styling can be observedand analysed

Several other sociolinguistic traditions, beyond variationism, arefully sensitive to contextualisation processes and have been so fromthe earliest days of sociolinguistics The ‘active contextualisation’perspective on social style that I am arguing for in this book is alreadyestablished in other parts of sociolinguistics, and was central to DellHymes, John Gumperz and others’ conception of the ethnography ofspeaking (Hymes 1962, 1996; Bauman and Sherzer 1989; Gumperzand Hymes1972) The theoretical tension that we have to deal with

in later chapters is in fact well summed up by the contrasting cations of the terms ‘speech’ and ‘speaking’ The variationist study ofsocial styles/ dialects has oriented to speech and to speech data, when

impli-it also needs to orient to speaking and to the styling of meaning insocial interaction This is not an oversight or even a limitation ofvariationist sociolinguistics in its own terms Variationism has simplyset itself other primary objectives, linked to understanding languagesystems and how they change, rather than understanding socialaction and interaction through language The objectifying priorities

of variationist sociolinguistics show through in much of its core minology The word ‘variation’ itself implies an analyst’s viewpoint,

looking down at arrays of variant forms distributed over some spatialmatrix What ‘varies’ is the community’s or the speaker’s languagesystem; more locally, what ‘vary’ are sociolinguistic variables (linguis-tic units of variable production) defined in the system This organisa-tion isn’t accessible to, or even directly relevant to, people engaged inspeaking and listening, although it is the variationist’s main concern.What matters to people is the meaning that language variation mightadd to their discursive practices – what people are trying to mean andwhat they hear others to be meaning.

Formal category systems and taxonomies used by researchers inmany fields of inquiry often imply equivalence between categorisedunits, along the lines of ‘this item is one of this type and goes here, andthat item is one of that type and goes there’ All research that is based

on coding and counting will make assumptions of this sort, andvariationist sociolinguistics does this too in some respects Variantforms of sociolinguistic variables are defined as being equivalent intheir referential meanings In the (English) Birmingham example, thephrase right time has the same linguistic (referential) meaning how-ever it is pronounced, and [aI] and [OI] are, to that extent, equivalent intheir meaning Whatever the speaker’s accent, the utterance seems toconvey the same basic information But this approach reduces thescope of the term ‘meaning’ and tends to wash out issues of value asthey attach to variable language in actual use When said in aBirmingham accent, the utterance and the speaker might conceivably

be held to be less convincing or authoritative, for example The socialmeaning of the utterance, depending on how it is phonologicallystyled, might interconnect in significant ways with other socialaspects of the speech event in which it is embedded

Bridging between survey orientations and practice orientations inthe sociolinguistics of variation seems an obvious development, eventhough the objectives and assumptions of (broadly) Labovian and(broadly) Hymesian sociolinguistics have traditionally been quite sepa-rate But the separation of these two agendas is in many ways artificial.There is a certain oddness in not addressing social interaction as amedium for variation research, in addition to its commitment tosocial surveying and to reaching generalisations at that level There

is no inherent clash between ‘macro’ and ‘micro’ levels of variationanalysis One important theme in later chapters is that local processes

of meaning-making depend on the affordances that socially tured variation in some sense provides, even though we need to befar more precise than this about how levels of analysis inter-relate.Speaking is the basic modality of language, where linguistic meaning

potential is realised and where social meanings of different sorts arecreatively implemented If we decide to engage with the idea of socialmeaning, however we precisely define it, social meaning will not besomething separate from the activation and interpretation of mean-ing in acts of speaking The term ‘discourse’ (despite the many differ-ent senses in which it can be used – see Jaworski and Coupland2007) is

a useful shorthand for this wider concern The research agendaaround style can therefore also be referred to as the analysis of ‘dialect

in discourse’

Quantitative analysis of the distribution of speech variants amonggroups of speakers is an abstraction away from the social process ofspeaking and of making meaning in context It is of course an entirelylegitimate research method, suited to its own purposes of generalisingabout language variation and change But investigating variation

in the context of social interaction is simply looking at languagevariation in its primary ecosystem of discursive meaning, and it cantherefore claim to be a sociolinguistic priority A more institutionalargument is that there should be benefits to any one tradition ofsociolinguistic research in reaching out to other traditions So much

of sociolinguistics nowadays is grounded in analyses of discourse andsocial interaction that, once again, it would be strange for variation-ism not to move into that arena This move might allow us to findother, more integrative, sorts of sociolinguistic truth

1 3 S T Y L E I N S O C I O L I N G U I S T I C S A N D I N S T Y L I S T I C S

It should already be obvious that the term ‘style’ has significant butlargely different histories in sociolinguistics and in other fields In thesociolinguistics of variation, style has been a very limited concept and

a peripheral concern In his overview of variationist sociolinguisticresearch Jack Chambers writes that ‘style is an important independentvariable but it is never the focal point (Chambers1995: 6) As we willsee in Chapter2, stylistic variation has been treated quantitatively insociolinguistic surveys in exactly the same way as social (or socialclass-related) variation is treated It has been a matter of demonstrat-ing that ‘intra-individual’ variation exists and that the nature of suchvariation can be explained by some simple principle or other In thissection, in order to gain some perspective, we return to some earlynon-sociolinguistic treatments of language style Naturally enough,there are many points of contact and overlap between early socio-linguistic treatments of style and early stylistics But those early

emphases and interests have in fact persisted much longer in tionist sociolinguistics than they have in stylistics itself Modernstylistics has blended into different forms of discourse analysis, pre-figuring some of the general arguments I am making in this book.The discipline label ‘stylistics’ was popularised in the 1950s, and itcame to be thought of as a discrete field of linguistics or appliedlinguistics ‘General stylistics’ (Sebeok 1960) was interested in allforms of language text, spoken and written, distinguished from thesub-field of literary stylistics Early stylistics was dominated by linguis-tic structuralism, which emphasised the structural properties of texts

varia-at different levels of linguistic organisvaria-ation (phonological, tical, lexical, prosodic) It gloried in the technical sophistication oflinguistic description, at a time when linguistics was still developingmomentum Stylistics was largely based on taxonomies – lists oflanguage features, levels and functions For example, a very simplehierarchical analysis of English style was offered by Martin Joos inhis strangely titled book, The Five Clocks (1962) The ‘clocks’ were levels

gramma-of formality in spoken and written English, which Joos labelled zen’, ‘formal’, ‘consultative’, ‘casual’ and ‘intimate’ It was based on

‘fro-an intuition about degrees of familiarity/intimacy between peoplewhich, Joos argued, impacted on communicative style The detail ofhow Joos meant these terms to be applied is not particularly impor-tant here, but the ‘clocks’ idea endorses a linear scale of ‘formality’.Formality or communicative ‘carefulness’ is assumed to dictate aspeaker’s stylistic choices or designs As we’ll see, this is how Labovcame to operationalise sociolinguistic style too

Roman Jakobson, in a famous lecture delivered in 1958 (Jakobson

1960, reprinted in Weber1996a), is often credited with giving the firstcoherent formulation of stylistics Jakobson’s theme was the relation-ship between poetics (aesthetic response to language and text) andlinguistics His argument was that the investigation of verbal art

or poetics is properly a sub-branch of linguistics He reached thisposition by establishing that the poetic function of language, which hedefined as ‘the set towards the MESSAGE, focus on the messageitself’ (Jakobson 1996/1960: 15; reprinted in Weber 1996a: 10–35), is ageneral function of all language use It is not restricted to poetry andother literary texts Jakobson argued that, if language always has apoetic function, linguistics must account for it, and that it couldand should therefore account for poetry and other artistic forms too.The most original aspect of Jakobson’s paper is his attempt to listall the main functions of language The poetic function stands along-side the referential function (the cognitive ordering of propositional

meaning) and the emotive function (affective and expressive meaning).Other functions are the conative function (organising meaning relative

to an addressee), the metalingual function (language ‘glossing’ or ring to itself) and the phatic function (language marking that people are

refer-in social contact)

This is a classically structuralist piece of theory, although Jakobson’sview of the multi-functional constitution of texts left a long legacy offunctional approaches to linguistics as well as to stylistics in particu-lar In the 1958 lecture Jakobson is in fact quite scathing about ‘thepoetic incompetence of some bigoted linguists’ (1996: 33), and atone point he quotes Martin Joos very disapprovingly for his excessivefaith in absolute categories For the contextual analysis of spokenstyle, Jakobson’s writing is in some ways liberating as well as inother ways constraining His claim that poetics deals with verbal

‘structure’ does seem to restrict the remit of stylistic inquiry to what

we can read from the surface of language texts – their linguistic forms

He gives us no hint that style has an interactional dimension, or thatstyling needs to be read and interpreted actively by listeners/readers.Similarly, he doesn’t assume that stylistic meaning is produced in theinterplay between textual and contextual processes, such as histories

of social relationships, ideologies of language or intertextual ships (echoes of meaning between different texts) His stylistics is tothat extent ‘technicist’ and formalist It puts too much emphasis onanalysts’ technical competence to reach analytic conclusions aboutstylistic effects It sees linguistic description as an analytic competenceand as a self-contained method

relation-There are echoes of these priorities in variationist sociolinguistics,particularly in its earliest detailing of stylistic variation in relation toaccent and dialect It would be surprising if this wasn’t the case, whenvariationism was being coined in broadly the same structuralist intel-lectual climate as early stylistics Even so, in Jakobson we also seeseeds of perspectives that came to challenge structuralist stylistics Hisset of linguistic functions already implied that language style involvedmeaning-making in different but simultaneously relevant dimensions

of a communicative act or event His ‘metalingual’ function days more commonly referred to as metalanguage and metapragmatics,see Jaworski, Coupland and Galasin´ski 2004; Richardson 2006)pointed to reflexive and self-referential processes at work in linguisticstyle He therefore opened a perspective on language in some waysreferring to itself, and speakers speaking through some level of aware-ness of their own stylistic operations and constructed images andidentities Although Jakobson’s main objective was possibly a rather

hegemonic one – to incorporate literary research into a fast-growinglinguistics – the way he foregrounded the poetic function of languageimplied that stylistics cannot ultimately be a purely descriptive exer-cise He showed that language styling, as I argued in section1.1, hascreative potential in the domain of aesthetics.

Linguistic function had been discussed much earlier, for example byKarl Bu¨hler (1934) It was Bu¨hler who first posited the functionalcategories of representational, conative and expressive (see Halliday

1996) Bronislaw Malinowski (1923) wrote about the phatic (ritualised,ceremonial) function of language (Coupland, Coupland and Robinson

1992), and J R Firth made important contributions to the development

of a theory of language genres, which he called ‘types of language’.But it was in Michael Halliday’s writing that the multi-functionality oflanguage was theorised in most detail (e.g Halliday1978) Hallidaymodelled linguistic meaning as being organised through threeconcurrent ‘macro-functions’, which he labelled ideational, inter-personal and textual These macro-functions could be followed throughfrom patterns of social organisation, with increasing detail anddelicacy, until they explained speakers’ lexico-grammatical and pho-nological choices at the level of individual utterances This isHalliday’s basic model of ‘meaning potential’ – what language canmean – and of language in use – how language means The model hasdeveloped into a general semantic theory of language called systemic-functional linguistics But it could also be applied, Halliday thought,specifically to the analysis of language style (Eggins and Martin1997,Leckie-Tarry1995)

Halliday introduced an abstract distinction between dialect andregister Dialect in this sense is language organised in relation to

‘who the speaker is’ in a regional or social sense, much as I introducedthe term earlier Register is language organised in relation to ‘what use

is being made of language’ Halliday treats register, or ‘languageaccording to use’, as a plane of semantic organisation, which can bespecified through the concepts of field (the organisation of ideationaland experiential meanings), mode (the organisation of textual andsequential meanings) and tenor (the organisation of interpersonalmeanings) So a particular register or way of speaking, if we treat it

as a uniform type or design of language use, will have distinctivesemantic qualities, reflecting speakers’ choices from the whole mean-ing potential of the language Ideational selections will show up astopics, things, facts or reports, most obviously in the grammaticalstructure of nominal groups Textual selections will relate to choices

of communicative mode/manner, sequencing, deixis and so on

Interpersonal selections will relate to social distance between kers, expressions of attitude, communicative ‘tone’ and so on Register

spea-or style, in Halliday’s conception, is the semantic spea-organisation oflinguistic choices taking account of communicative purposes andcircumstances (see the useful review in Gregory and Carol1978)

In Halliday’s functional linguistics we see style emerging from themargins of linguistic theory and description, and being highlighted as

an inherent dimension or set of dimensions of language organisation.Style is an inherent part of all communicative activity Halliday says it

is wrong to equate style with ‘expressive’ function alone:

Even if we are on our guard against the implication that the regions oflanguage in which style resides are linguistically non-significant, weare still drawing the wrong line There are no regions of language inwhich style does not reside (Halliday1996[originally 1965]: 63)

He resists ‘an unreal distinction between the ‘‘what’’ and the ‘‘how’’ and how they may be incorporated into the linguistic study of style’(1996: 64) Register is as much about the ‘what’ of language use, such

as what gets talked about and in what terms, as it is about the ‘how’ oflanguage use There is no act of speaking without a register or styledimension at work within it

As a theorist of grammar and meaning, Halliday has mainly beeninterested in explaining the organisation of language texts Systemicfunctional grammar (as Halliday’s theoretical approach is known) islargely an attempt to model the increasingly detailed meaning choicesthat speakers make, and how meanings come to be realised in partic-ular utterances It is in some ways a sociolinguistic theory of language,because it tries to trace meaning choices that are made available inparticular social contexts It is of course true that many meaningchoices in discourse reflect the social context of speaking in a ratherdirect and simple way Speakers, for example, use technical vocabula-ries associated with specialist topics, purposes and ‘registers’ Onmany other occasions, the link between context and style is far lessdirect, less determined and more subject to speakers’ and listeners’creative agency Systemic functional linguistics has not specialised inmodelling variable language use at the level of accents and dialects.But its general perspectives on style are useful for sociolinguistics,particularly in stipulating that style is socio-semantically motivated Itemphasises that style is part of the process of meaning-making indiscourse

Halliday’s concept of register grew out of a theoretical tradition,mainly in British linguistics, that had for some time been interested in

the link between language use and social situations J R Firth (1957)coined the phrase context of situation, pointing to local (objective andsubjective) norms that constrain linguistic style, as in the simple andoften-repeated instance when the environment of a church or mosquemight be linked to silence or whispered talk Firth twinned the termcontext of situation with the term context of culture, suggesting anested arrangement of stylistic constraints A culture defines a con-text for social interaction at a macro level, which is then specified intodifferent social situations.

Despite Halliday’s argument that style and register imply a tous dimension of complex meaning organisation in texts, the term

ubiqui-‘register’ has usually suggested a fixed relationship between ‘a style’and ‘a social situation’ An example would be the idea that newsreading on television would be delivered in a register or style ofnews reading The idea is obviously trite, although it captures a gen-eralisation of sorts about social styles – there certainly are stylistictendencies in ‘news reader speak’, even though they would not beunique to news reading, and listing them would be a rather tedioustaxonomic exercise The theoretical limitation is that, if a register isdefined by the situation that it accompanies, there is no linguisticwork for the concept to do As Judith Irvine implies (Irvine2001: 27),defining registers, and therefore styles, as situational varieties mayhave resulted unfortunately from Halliday’s theoretical twinning ofdialect and register If dialects are presumed to be discrete regionalvarieties of a language, then perhaps it seems reasonable to presumethat registers are discrete situational varieties of a language Irvineagain makes the point that social situations are in fact often distin-guished by types of speakers populating them, and vice versa, so eachdimension implies variation in the other In fairness, this is just whatHalliday stressed – that dialect and register needed to be seen as twosides of the same coin, and not as independent dimensions of linguisticorganisation and difference Even so, the ‘twin dimensions’ approach

to dialect and register lowers our analytic expectations in relation toeach It endorses the view that variation can be explained in linearterms, and it points us to simple sets of categories in each dimension.The variationist model maintains these same assumptions

In fact, the statistical and correlational linking of speech style andsocial situation has lived on in variationist sociolinguistics, wherestylistic stratification is defined as speakers speaking differently indifferent situations (see Chapter2) In sociolinguistic interviews thephysical situation does not change, but types of speech activity aremanufactured to introduce different levels of attention to speech

by interviewees, and hence different ‘situations’, subjectively enced In this way of thinking, speech style is predicted on the basis ofboth dialect (relating to who the speaker socially is) and register (whatsituational constraints are operative), together ‘Speaking differently’

experi-is measured by the quantitative means mentioned in section 1.2 –based on how frequently particular speech variants are used by spea-kers It is worth stressing again that this offers a statistical definition

of ‘a style’ or ‘a stylistic level’ A particular speaker taking part in adefined speaking activity is said to be using or producing a ‘style’which is actually a numerical index of ‘overall degree of standardness’

on an abstract scale

While the concept of register has not found much favour in porary sociolinguistics, the concept of genre is very firmly established(Bakhtin1986, Macaulay2001, Swales1990) Common definitions ofgenre tell us that genres are culturally recognised, patterned ways

contem-of speaking, or structured cognitive frameworks for engaging in course So the most clear-cut instances are institutionalised communi-cative genres, such as political speeches, lectures, post-match sportsinterviews or stand-up comedy routines In these cases quite specificframeworks exist, and indeed there are often partial scripts, for how

dis-to fill out the discourse of a genre People recognise these genreswhen they come across them, and they can refer to them throughfairly simple labels; they appreciate their norms and their discursivedemands on people taking part Once again, this fits into a generaldefinition of social styles Our socialisation into a cultural group’s ways

of communicating is partly a matter of learning institutional genres –learning how to ‘read’ them and sometimes learning how to enactthem, and coming to appreciate their social resonances and values.Other genres are much more diffuse Should we, for example, con-sider conversation to be a genre, or is better to think of sub-types ofconversation as genres? Is banter a genre, or small talk, or gossip, orverbal play, or argument, or flirting, or story-telling (J Coupland

2000)? Or is ‘flirtatious verbal play’ a genre in itself, and is ‘gossipystory-telling’ another one? But even diffuse, less institutionalised andhard-to-label ways of speaking like these will meet the main criterionfor genre This is the criterion that participants have some significantawareness, as part of their cultural and communicative competence,

of how the event-types they are engaging with are socially constituted

as ways of speaking They will, at least to some extent, appreciate theconstraints and opportunities that a particular genre brings with it

We will need to build on this core idea of genre in later chapters Forthe moment it is enough to emphasise some fundamental points

about the relationship between a notion of genre and early notions

of style

First, when we think of speech genres, we are pulling together whatotherwise seem to be different levels of ‘the social’ – cultural salienceand local acts of speaking This is one way in which there is a necessarylink between the local organisation of talk and macro-level socialstructure To understand speaking and styling as sociolinguistic pro-cesses, we have to entertain a notion of social organisation that bringstogether situational and cultural contexts, much as J R Firth hadoriginally suggested Second, any notion of genre is an interactionalnotion – it specifies social positions, roles and responsibilities forsocial actors, and usually multiple participants In the conceptions

of style that have come up so far, we have mainly been concerned withthe talk or text produced by a single person, or people groupedtogether in abstract ways (recall the ‘young Birmingham women’),speaking under certain conditions (recall ‘formal talk in an inter-view’) Third, genre gives an idea of social context where it is clearthat the organisation in question is partly pre-figured in the socialenvironment (culturally recognised and endorsed) and partly con-structed by speakers themselves When we embark on a sequence ofgossip, we have an initial understanding that conversation can go thisway, that there are specific possibilities and sanctions attached to it,and perhaps that there are specific costs and outcomes But there is nocontextual ‘flag’ signalling that this is how we must now converse.Gossip is often initiated through some subtle process of discursivenegotiation whose result may be some sort of consensus that ‘we arenow gossiping’

Variationist sociolinguistics, and studies of style in that tradition,have very rarely entertained any notion of genre, although it is afundamental concept for the analysis of social meaning I will bearguing later on that even the social meaning of particular socio-linguistic variants depends on a reading of genre and social context inthat sense In fact we will have to go much further into how socialcontexts are constituted than just asking ‘what genre is this?’ Styling

is part of the process of genre-making, but also part of the process ofgenre-breaking Styling can reshape conventional speech genres andhow we expect to participate in them For this sort of analysis we willalso have to engage with theoretical ideas like discursive frame anddiscursive stance, which describe perceived qualities of social interac-tion operating more locally than genre

As I have already mentioned, some sociolinguistic theories of socialcontext have been concerned with active, local meaning-making

for a long time In these approaches it makes less sense to talk about

‘styles in contexts’ and more sense to talk about processes of isation – sociolinguistic style creating context as well as responding tocontext Arguments of this sort were made as far back as the 1970swhen, for example, social psychologists of language argued thatspeech style should not be approached in a ‘static’ way but in a

contextual-‘dynamic’ way (Giles and Powesland1975) Social psychologists werepredisposed to seeing contexts as the outcome of subjective processes

As Howard Giles says in a reflective comment about earlier research ofhis own with colleagues, they were interested in ‘how speaker-hearerscarve up contexts psychologically and subjectively’ (2001: 211) Thisidea opens up important possibilities, and not least that differentpeople might construe any given social context differently, with theimportant implication that ‘the current context’ (or genre or frame,and so on) often has a degree of indeterminacy about it Context is alsoamenable to tactical manipulation, and one participant can engineeranother’s understanding of ‘where we stand in this context’, perhaps

to shock, to amuse or to confuse This is one of those circumstanceswhere Jakobson’s metalinguistic function comes into discursive play.The social psychology of situational construals has not typicallytracked local contextual manipulations of this sort, even though

it has provided the conceptual apparatus to do this sort of analysisfor a long time

In sociolinguistics, John Gumperz developed a view of active text formation through his notion of contextualisation cues (Gumperz

con-1982: 130–52) In conversation, speakers routinely signal to othershow aspects of what they are saying should be heard and analysed

A discourse marker such as oh at the head of an utterance, said with

a short falling intonation from a high start, perhaps accompanied by araised eyebrow, can signal that what follows is likely to be a disagree-ment Gumperz says that contextualisation cues are links betweensurface style features and how the content of talk is to be understoodand ‘what the activity is’ (1982: 131) But cues can create contexts inother, less consensual ways too Accent features, intonation featuresand so on can lead to inferences, correct or not, about a speaker’ssocial origins or communicative competence Gumperz calls this pro-cess conversational inferencing and shows that it is a potentially damagingprocess of social labelling and attribution Social attributions – forexample associating forms of speech with gender, age, class or racialcategories, and inferring competences of personality characteristics

to them in turn – are made possible through social stereotypes(Hewstone and Giles 1986) Once again we see a theoretical nexus

between local happenings in talk and socially structured beliefs andexpectations, and this is the territory in which sociolinguistic stylingoperates.

This introductory discussion shows that issues of social context are

at the heart of any analysis of language style, but also that there aremany different ways in which sociolinguistics can address social con-text The main distinction is between approaches that pre-determinecontext, recognising or even consciously setting up ‘social contexts’within which to analyse style variation, and approaches that invertthis relationship In that alternative perspective, style lives in a dia-logic relationship with context Context (as in the concept of genre) is

in part a socially structured phenomenon that speakers have to scribe to and that they often live out in their talk But context is also, inpart, the product of their discursive operations Variationist socio-linguistics has stuck with deterministic formulations of context andnot generally explored the implications of social construction As weshall see, that constructionist impetus has come more from anthro-pological linguistics and discourse analysis The active/verbal/agentivesense of the term ‘social meaning’ becomes important in a construc-tionist analysis, and we need to review this core concept further, inthenext section

sub-1 4 S O C I A L M E A N I N G

Social meaning has always been a relevant concern in sociolinguistics,but what exactly does it refer to? Sociolinguistics is an exploration of

‘the social significance of language’, although we can unpack this idea

in different ways Linguists might assume that the domain of meaningbelongs to them, but in fact social meaning is a core concern of manydisciplines It can refer to how we impute meaning to, and take mean-ing from, our cultures, our communities, our personal histories, oursocial institutions and our social relationships Cultural values andnorms, social power and status, intimacy and distance are all socialmeanings Then there are the meanings we invest in our own andother people’s social positions and attributes – selfhood, personal andsocial identities, social stereotypes, prejudices, conflicts and bounda-ries These concepts already go a long way towards defining the prob-lems and questions of all forms of social science, sociolinguisticsincluded

Many of the social sciences are interested in social meaning in alinguistic sense too, because they recognise that language provides

the salient fields of action for so much of social and cultural life.

A large slice of contemporary sociology, anthropology, social ogy, communication/media studies and other related disciplines isavowedly ‘discursive’ (and this usually means taking a ‘social con-structionist’ view of the role of language) These disciplines generallyrecognise the constitutive power of language in the structuring of socialcategories and social life in general Discursive/interactional socio-linguistics shares these assumptions too Language-based disciplinesare generally better equipped than others to undertake analysis ofsocial meaning when there is an explicitly linguistic analytic focus,but this potential isn’t always realised A first step might be for socio-linguistics to widen its own remit when it comes to social meaning.The range of issues I have just sketched out is massive and daunting

psychol-I am not suggesting that an interactional sociolinguistic approach can doadequate service to all of them But there is a stark contrast betweenthe narrow sense of social meaning that has dominated in variationistsociolinguistics and the extremely broad reach of the concept else-where A social constructionist approach to social meaning cannotavoid reaching into complex territories of cultural, personal, histor-ical and sequential meanings This is its strength and its weakness But

I will be arguing that sociolinguists should go after this sort of ity of social interpretation, simply because social interaction itselfimplicates this level of complexity

complex-It is useful to look at an influential and representative variationistsociolinguistic view of social meaning Jack Chambers writes that ‘themost productive studies in the four decades of sociolinguistic researchhave emanated from determining the social evaluation of linguisticvariants’ (Chambers2004: 3) I am sure Chambers is using the phrase

‘social evaluation’ as a synonym for ‘social meaning’, even though itmight be preferable keep the term ‘social evaluation’ for the process

of judging speech varieties or speakers Social meaning is at the core oflanguage variation research because, as Chambers says, ‘the variantsthat occur in everyday speech are linguistically insignificant butsocially significant’ (2004: 3) His examples in an introductory discus-sion are these:

Adonis saw himself in the mirror

Adonis seen hisself in the mirror

These examples, and particularly the second one, may not strike us as

‘everyday’ utterances, even in the hypothetical context, as it might be,

of someone talking about a scene they have just watched in a classicalplay But the examples clearly make the point that different linguistic

forms can express what is referentially the same meaning, whiledifferent social nuances are present (This is broadly the definition of

a sociolinguistic variable and its variant forms that we discussed insection1.2.) Chambers is pointing out that the grammatical meaning

of past tense in English can be expressed either by saw, which isconventionally called ‘standard’ linguistic usage, or by seen, described

as ‘non-standard’ usage Different forms of the reflexive pronoun –himself versus hisself – stand in the same relationship to each other Thelinguistic or referential meaning is unchanged whichever form isspoken, but, Chambers points out, the sentences ‘convey very differ-ent social meanings [and] sociolinguistic significance’ (2004: 4).The second sentence of the contrasting pair is rather captivating,socially and contextually, if we try to analyse it as an act of speakingrather than just as a constructed example of ‘non-standard’ language.Who could have said this, in what circumstances and why, and whatsocial meaning would we impute to the speaker or to the socialarrangements that might have made this a sayable and interpretableutterance? The most striking aspect is the wonderful mis-match – thesemantic dissonance – between the utterance’s referential or idea-tional meanings (what its words denote or refer to) and the vernacular(or ‘non-standard’) dialect forms We have the classical, mythological,high-culture moment of Adonis seeing his own image in a mirror,voiced through vernacular English dialect grammar Past-tense seenand the reflexive pronoun hisself are certainly English dialect forms wecan find in common use (see, for example, Cheshire 1998) But astylistic sociolinguistic analysis (rather than a dialectological analysis)would point to a clash of stereotyped social milieux, not to simple

‘variation’ Jenny Cheshire writes about ‘non-standard’ grammarusing features of this sort among young people in Reading (a city tothe west of London) indexing ‘vernacular culture’, and (notwithstand-ing the fact that the language code is English) Adonis and his mirrorcan be assumed to reside in a different cultural field So we mightreach for social explanations in terms of genre or register Is theutterance Adonis seen hisself in the mirror said in parody? Or might itactually be a moment from classroom discourse in Reading wheresome school kids have been required to sum up the action of a play?

Is it a studiedly anomalous bit of meaning-making of the sort thatattracts attention and humour? (I suspect we have hit upon JackChambers’s own motivation.)

Chambers is certainly not seeking to make points like these Hisobjective is simply to introduce the concept of social meaning in adiscussion of language variation But I think his discussion of these

examples hints at what does and does not generally matter aboutsocial meaning to variationists Chambers says that the first sentence

‘is emblematic of middle-class, educated or relatively formal speech,while the second is emblematic of working-class, uneducated orhighly colloquial (vernacular) speech’ (2004: 4) Firstly, this viewassumes that a direct indexical relationship exists between a socio-linguistic variant and a social meaning And secondly, it readssocial meaning mainly in terms of social group membership andsocial identity in that category-bound sense Chambers uses the idea

of ‘emblematic’ status to express the direct link between grammatical

‘standardness’ (my scare-quotes) and ‘middle-class, educated or tively formal speech’ (Chambers’s words) (In fact, he might mean thatgrammatical ‘standardness’ is emblematic of middle-class-ness andeducated-ness, because grammatical ‘standardness’ is a way of speak-ing, not an emblem of a way of speaking.) The main assumption here

rela-is that the grammatical ‘standardness’ of past tense saw stands for(connotes, implicates, signals, evokes, indexes) being a member of thesocial group we know as ‘educated middle-class’, and so on Later inthe same source Chambers in fact says that the social significance oflinguistic variants is very often not an attribute of their presence orabsence in a person’s speech, but rather of their frequency in thatperson’s speech compared to someone else’s speech (2004: 115) But

we can come back to the complicating issue of statistical frequenciesand their connection to social meaning in Chapter2

So does ‘standard grammar’ – always and necessarily – cally signal that a speaker is a member of the ‘educated middle-class’?Apart from the severe difficulties of defining these sociological terms,

emblemati-we have not yet taken account of contextualisation A key problemwith the terms ‘standard’ and ‘non-standard’ (and one of my reasonsfor scare-quoting them) is that we can really only understand one ofthem in relation to the other There might be some social shadow

of ‘educatedness’ around ‘standard’ grammar if we are made aware ofthere being a shadow of ‘uneducatedness’ around a ‘non-standard’

or vernacular alternative grammatical form The social meaning thatChambers posits seems to be an effect of putting the two utterancesnext to each other as examples, more than a result of the inherently

‘emblematic’ status of either Following Irvine (2001), Ben Ramptonstresses ‘the indelible relationality of styles’ (2006: 379, note 5) This isthe same point that I made earlier, that styles achieve their meaningthrough contrast and difference If we take the view that ‘standard’grammar is ‘least exceptional’ language – a moot point, but Adonis seenhisself in the mirror isn’t an unexceptional utterance form – then the

grammar of the ‘standard’ equivalent sentence isn’t truly emblematic

of anything; it is unexceptional

There is obviously some need for further clarity in this sort ofdiscussion, and it will be useful to look back to early semiotic theorywhere concepts in the general area of ‘standing for’ relationships werefirst developed The link between an expression or form and what itmeaningfully stands for is usually referred to by the term indexicalitythat has already come up in the discussion (but see also Milroy2004,Silverstein1976) The formal definition of an index was conceived byCharles Peirce (1931–58) Peirce said that an index is a relationshipbetween a sign and a referent (the object that it is linked to) which isbased on a physical or in some other way objective or ‘real’ associa-tion For example, a bullet hole ‘indexes’ the fact that a bullet haspenetrated a surface An index can in theory be distinguished from anicon, which is where we perceive some sort of natural resemblancebetween the sign and the object that it signifies, such as when aphotograph provides an iconic ‘likeness’ of a person A third type ofrelationship occurs with symbols, where societies forge links that areoriginally arbitrary between signs and meanings, such as an eaglebeing taken to stand symbolically for authority Using these termscarefully, following Peirce, we would have to say that a grammatically

‘standard’ variant, treated as a sign in semiotic theory, has symbolicmeaning, because the link between it and being middle-class is arbi-trary rather than natural or objective

The study of language ideologies – the study of how languages andlinguistic styles or features come to have given social and ideologicalmeanings – suggests ways in which links of this sort can in fact bereshaped (Gal and Irvine1995, Irvine2001) There is the process ofnaturalisation, when arbitrary signs that we would technically callsymbols are treated as if they were (natural) icons or (objective)indexes We can see that technically arbitrary or meaningless bits ofsound and linguistic form, like features of accent and dialect, veryfrequently come to have indexical-type meaning People come tobelieve that using a particular accent caries the ‘objective’ or ‘natural’meaning of ‘low social class’ or ‘uneducated speaker’ The processcalled recursion refers to the expanding of a meaning relation, forexample when the meaning ‘uneducated speaker’ gets attached to asingle speech feature This might be the grammatical feature ‘non-standard past-tense seen’ as in Chambers’s example, or the phonolo-gical feature often referred to as ‘G dropping’ – using alveolar [n]instead of velar [N] in words like waiting, seeing, something and nothing.The ideological process of erasure is when a pattern of meaning

associations is simplified, and one part of the meaning complex isforgotten about or elided (Gal and Irvine1995, Manning2004).These ideas about language-ideological processes help us to see thatindexical relationships (using this now as a general term) are notentirely stable over time Recursion and erasure might come aboutthrough slow historical processes of change, as the social meanings of

a linguistic form or pattern gradually shift But it is quite feasiblefor speakers to bring about similar shifts locally in their talk Theycan, for example, creatively forge a new association between a linguis-tic form and an individual or group not previously linked to it Othersorts of shift are also possible Penelope Eckert writes about theprocess of stylistic objectification in young people’s social development:social development involves a process of objectification, as one comes

to see oneself as having value in a marketplace (Eckert2000: 14).She argues that ‘at this point, speakers can point to social meaning –they can identify others as jocks or burnouts [group labels that youngpeople use to mark their pro-school or anti-school orientations – seesection2.5], as elite or working-class, educated or not, prissy or tough’(2000: 43) Therefore, sociolinguistic indexicalities are sometimesmatters of social attribution, and they become amenable to being dis-cussed, argued over and renegotiated, metalinguistically

So, even when we are dealing with social meaning in terms of theindexical potential of social styles such as accent/dialect features,individually or in bunches, we have to be aware of complexities andpossible instabilities in meaning relationships We should not expectlinguistic features to have unique social meanings, even in the samesocio-cultural settings Scott Fabius Kiesling’s (1998) study of the com-plex social meanings of the ‘G-dropping’ variable used by young men

in college fraternities in the USA is an excellent case in point Kieslingsuggests that the ‘non-standard’ [n] form of the (ing) variable amongmale students can index the social attribute of being ‘hardworking’, orhaving a ‘casual’ approach or being ‘confrontational’ In Kiesling’sview the [n] feature has no meaning as such, and acquires meaning

‘only when an identity takes shape through the tension between thetext and content and the negotiation between speaker and hearer’(Kiesling1998: 94)

In later discussions we will need to return to these active tualisation approaches to indexicality They clearly undermine theassumption of ‘one form, one social meaning’ But they also implythat we should look for social meaning in different places It is notthe case – or at least it is not only the case – that language forms are

allocated meanings by the sociolinguistic system and then ‘selected’locally We will need to think in terms of social meaning potential (touse Halliday’s phrase) being called up or activated or validated, orundermined or challenged or parodied, in particular discursiveframes for particular local effects This would imply, once again,that social meaning doesn’t exclusively reside in linguistic forms, oreven in so-called speech communities or in speakers’ sociolinguistichistories and experiences It is partly a situated achievement in acts

of observing speech would make a speaker speak self-consciously andtherefore unnaturally This was the basis of his method for elicitingstyle-shifts in interviews (see Chapter2) This line of argument repre-sents communicative reflexivity as a methodological problem

The standard procedures of the sociolinguistic survey interviewwere developed as a way of getting round the apparent problem

of the observer’s paradox Observation remains the key method forvariationist sociolinguistics, and sociolinguists often feel that need

to ‘leave the laboratory’ and ‘get out there’ into the ‘real world’ oflanguage use Crawford Feagin (2004), for example, writes aboutthe need to ‘enter the community’ to solve the observer’s paradox

Technical research apparatus to do with sampling, recording, scription and formal analysis follows on from this A concept of ‘gooddata’ exists in variationist sociolinguistic surveying and it relates tocriteria of naturalness, untaintedness and representativeness, as well

tran-as to the need to get excellent acoustic quality in audio-recording.These priorities follow from the primary objective of discoveringhow linguistic systems are structured and how they might bechanging

Any study of speech style, including research targeted at linguisticvariation’s role in the construction of social meaning, has to engagewith these classical problems of sociolinguistic method All socio-linguistic studies need ‘good data’, even though they will interpretthis idea differently But the empiricist assumptions driving socio-linguistic observations introduce their own problems of theory andmethod One of these problems is the basic assumption that speak-ing is ‘real’ or ‘natural’, provided it is not observed As we will see, it

is well worth exploring what lies behind these assumptions andbehind the general appeal to ‘sociolinguistic authenticities’ and

‘the authentic speaker’(see section7.2) Is it in fact possible to definenaturalness in speaking, and to determine when speaking is and isnot natural? Is it enough to rely on sampling procedures and cleverdevices in the design of interviews to gain access to the ‘ordinary’ or

‘everyday’ usage that variationists value? What is ‘authentic speech’and what defines authentic speaking (Bucholtz 2003, Coupland

2003, Eckert2003)?

As we will see later, these may not even be the most profitablequestions to ask Instead of either glorifying authenticity or dismiss-ing it out of hand, we can approach it in other ways Authenticitycould be a powerful concept to use within the analysis of style Styling,for example, creates social meanings around personal authenticityand inauthenticity, when speakers parody themselves or presentthemselves as ‘not being themselves’ Erving Goffman (e.g.1981) hasgiven us intriguing insights into how performance and theatricalityintrude into everyday social practices, and sociolinguistic variationgives us resources to ‘stage’ our identities in many different ways Wecan think about ‘self-authentication’ and ‘other-authentication’, butalso ‘de-authentication’, as strategic possibilities for how we constructidentities in talk

The conventional wisdom around authenticity has been far morestraightforward Sociolinguistic surveys have tended to assume thatspeakers are, in themselves, authentic members of the groups and the

‘speech communities’ that they inhabit – recall our Birmingham